The International Prize Pool Down 91 Percent and the Global Reallocation of Esports Capital

Every great spreadsheet begins with an empty cell and a question.

This time the empty cell sat in the "salary obligations" column of my tracking file, dated September 6, 2026. I filled it with one line: 3 billion won. Roughly USD 2 million, for Dplus KIA's League of Legends roster alone. Days earlier, that team had won the League of Legends title at the Esports World Cup 2026. Reporting from Korea then said the organisation had delayed player salary payments and was seeking a new owner.

The trophy and the cash flow sit in the same spreadsheet. They do not share a status row.

Not long before, in a different title, Falcons — the team that won The International 2026 in Dota 2 — announced it was withdrawing from Dota 2. During 2026 the organisation entered 18 tournaments across the Esports World Cup system. It won the biggest event in a title, then left that title. Falcons' official statement referred to "long-term sustainable operations".

Two champions. Two withdrawals nobody drew on paper in advance. I reopened my The International prize-pool file and found a curve that had already finished drawing itself.

A method note before going further

Of the entire dataset behind this piece, only one statement is attributed to a named source: the Falcons statement. Everything else is published data or commentary explicitly labelled as the author's assessment. I say this first because the argument below rests on the quality of the columns, not on how smoothly the sentences run.

One marker matters. The The International prize-pool figures for 2026–2026 — USD 40 million, USD 18.9 million, roughly USD 3.4 million — match the tournament's public record. That overlap gives the surrounding data an initial credibility reading. Every other figure here should be treated as pending independent verification.

Context: one gear pulled out of the funding engine

Dota 2 once ran on a mechanism with almost no precedent in professional sport. Valve sold in-client items through the Battle Pass, and a share of that revenue went straight into The International prize pool. Fans bought items and, in doing so, they directly paid the year's biggest tournament. The prize pool became a measure of community participation rather than a publisher budget line.

The model peaked in 2026 at about USD 40 million, a level no esports event had reached. In 2026 the pool was USD 18.9 million. In 2026, about USD 3.4 million. In recent editions, low single-digit millions. Against the peak, a decline of roughly 91 percent.

That curve does not describe a community turning away. Valve reworked the Battle Pass structure and severed the link between item sales and the prize pool. The gear came out of the machine. The International's prize money moved from a pool the community filled to a sum the publisher decides.

When I watched the 2026 broadcast from Seoul, viewers in the room talked about Battle Pass items almost as much as about team fights. The prize pool rose while the audience watched. When the mechanism changed, the attention stayed; the money simply stopped travelling along the same line. The correlation between the two was once very strong. The causation always sat on the product side.

The opposite pole: capital from a different origin

Alongside Dota 2's falling pool, another tournament structure grew. The Esports World Cup 2026 distributed USD 75 million across dozens of titles. The Saudi eLeague 2026 recorded more than SAR 4 million and 37 clubs. This is state-linked, multi-title capital that does not depend on a single game.

In Korea, the LCK imposed a salary cap with a luxury tax. That is a league-level intervention aimed at competitive balance and long-term viability — a sharing mechanism in which the highest-spending organisations contribute back to the rest of the league.

Three pieces — a publisher-determined prize pool collapsing, Gulf capital expanding, a Korean league applying its own brakes — form a picture I read as reallocation, not recession.

The evidence chain

Six rows carry the argument. First, the The International curve: USD 40 million in 2026, USD 18.9 million in 2026, about USD 3.4 million in 2026, low millions recently, a roughly 91 percent fall from peak. Second, Valve's Battle Pass rework — a product-layer change, not a gameplay-balance change. Third, the Esports World Cup 2026 at USD 75 million across dozens of titles. Fourth, the Saudi eLeague 2026 with over SAR 4 million and 37 clubs. Fifth, Falcons: The International 2026 Dota 2 champions, 18 entries in the Esports World Cup 2026 system, then withdrawal from Dota 2. Sixth, Dplus KIA: Esports World Cup 2026 League of Legends champions, a roster costing about 3 billion won, delayed salaries, and a search for a new owner.

Read together, the central conclusion is this: competitive strength and financial survival have decoupled. The industry's implicit assumption used to be that winning saves you. A major title brought sponsors, new contracts, and enough cash flow to cover payroll. That assumption has now been rejected twice in a single cycle, in two different titles, by two organisations that both won.

The second conclusion is harder to hear: the current problem in esports is distribution, not the total volume of capital. The money still exists inside the system. It just no longer flows evenly through every layer. Capital clusters around major tournaments, around titles with clear commercial value, and around organisations with sustainable operating structures. The rest must improvise.

The mechanism behind that clustering is a race in which salaries outrun revenue. During the growth phase, player prices climbed faster than organisations generated income. A roster worth millions that produces little matching commercial value becomes a liability on the balance sheet rather than an asset. That is exactly Dplus KIA's position: an LoL roster consuming about 3 billion won while the organisation's revenue side failed to keep pace. The LCK salary cap arrived as a necessary correction, not a punishment.

Falcons has to be read through a different logic. This is not a competitive failure. An organisation that wins The International 2026, enters 18 events inside one system, and then exits a single title is performing portfolio optimisation. It is reallocating away from a lower expected margin and toward better commercial and geopolitical leverage. Falcons keeping many other titles confirms that reading. The word "withdrawal" changes meaning: no longer a signal of collapse, but a signal of reallocation.

Taken together, the polarity is visible. One pole holds major tournaments, Gulf capital, and multi-title organisations with disciplined cost structures. The other is a long tail of single-title, prize-money-dependent organisations with payrolls inflated during the growth years. The two poles do not move in the same direction, and their divergence is the real content of this period.

At the tournament-system level, the structure of power is shifting too. Prize money has moved from community crowdfunding to publisher discretion, which means a single product decision can reshape an entire title's economy. On the other side, third-party events backed by state capital own the organising rights but not the game rights. These two forms of power have no reconciliation mechanism, and that gap is the largest thing my data cannot fill.

The contrarian angle

The most common reading of Dota 2's prize curve is "esports is dying". That reading fails at the arithmetic level. The fall from USD 40 million to a few million is a direct consequence of removing crowdfunding, not a measure of player interest. Merging the two is an error the data here does not permit.

The second reading, which sounds more sober, is "it is only reallocation, nothing to worry about". That is wrong in the other direction. Reallocation always has a payer, and here the payment is made in delayed contracts, dissolved rosters, and organisations leaving the title they once won. A process described with a neutral word can still cause non-neutral losses.

The most important contrarian point is this: the party that decides a title's fate may not be the party that competes — it may be the party that sells the items. Valve changed the Battle Pass structure and a sponsorship channel worth tens of millions of dollars vanished. No analysis of that change's effect on competitive balance accompanied it. No cross-publisher safeguard exists. This operates exactly like an invisible referee: a patch can decide a championship, and here a product decision can decide whether an ecosystem survives. We usually call that "meta adaptation". In reality, most of the adaptability credited to organisations is simply the result of standing on the right side of a capital flow they do not control.

Alternative hypotheses deserve space. The low The International pool may not signal Dota 2's weakness but a deliberate Valve strategy — moving from publicly tracked prize-pool racing toward in-client monetisation. If so, the curve measures a change of objective, not a change of health. Second: Gulf capital is a subsidy masking a structural deficit elsewhere, and when that subsidy plateaus, contraction will be faster than expected. Third, less likely: Dplus KIA's owner search is an ordinary asset sale rather than a distressed disposal. The cash-flow evidence leans the other way, but I cannot fully exclude it.

And one limit cannot be skipped: sample size. Two organisations, one publisher's product decision, one regional cap. That is a small sample for a large conclusion. Error does not lie — it only whispers what we are not yet large enough to hear.

What stands out is the near-absence of the rest of the world. China, Europe and North America do not appear in the dataset I hold. An analysis of a global ecosystem missing those three regions has its blind spot in its own structure, not in its conclusion.

What to watch in the next cycle

I am not predicting whether esports rises or falls. That prediction needs a variable nobody has published. I am holding four verifiable signals for the coming months.

First, whether the LCK salary cap spreads to other leagues. If it does not, Korea faces a new balance problem: keeping stars inside a spending-limited league while uncapped leagues keep their doors open. Second, whether Falcons returns to Dota 2. A return would refute the portfolio-optimisation reading; prolonged silence would confirm it. Third, whether appearance fees become the base income for mid-tier organisations. If they do, those organisations' revenue depends on being invited rather than on winning — a new category of risk that has not yet been named. Fourth, whether another publisher unilaterally restructures its own title's funding channel. If so, the sample grows and the conclusion above shifts from hypothesis to rule.

The transfer market is where emotion gets beaten by probability. In this window, probability leans toward organisations with multiple titles, disciplined costs, and a place inside a stable capital flow. Everyone else will choose between shrinking and disappearing.

A shock is only data that history has not yet read by name. Two championship trophies failed to save two organisations, and that is the first data row of a new chapter. On the night itself, the biggest tournament's lights stayed on, the stands stayed full, and there was simply one fewer jersey on the bench. Nobody in the arena noticed. I noticed the next morning, opening the spreadsheet and finding an empty row in the budget column where a champion's name had been a year earlier.